twitter, FX and working in Europe

Working in Europe got me thinking about my currency exchange timings and so have been working on the analysis of GBPEUR, GBPUSD, EURUSD and the reciprocals, USDEUR, USDGBP and EURGBP and these are now being tweeted on @space_FX or https:twitter.com/space_FX.

The London Stock Exchange twitter feed name has changed to space_LSE - @space_LSE or https://twitter.com/space_LSE

But back now and have been trading here on the space_LSE feed with some very good results.

ORE II

its taken a little more than the 5 days, and this will be logged in the current Markov Chains, but ORE has now reached its desired level and continuing up. :-)

And legs for opera

As you will probably have seen, Twitter had changed its authorization scheme from XML to JSON meaning that the Oracle PLSQL XML based tweets no longer worked. I have been working hard to re-establish this process and have now completed my testing, and we are tweeting from oracle once again.

I have just Tweeted all companies that have been missing since Twitter changed the Authorization on the 11th June 2013. The twitter link is in the heading.

I have been able to accomplish v1.1 twitter app authorization by writing an external Perl script. This is not as difficult as it first appears although you will need to get all your ducks lined up in a nice row.

Chronologically, this is done by:

1. go onto http://search.cpan.org/ and first off make sure that your Module::Build is up to date - there are a couple of ways to do this dependent on if you are working with windows or linux. linux is pretty easy because one runs cpan as root, and then types in 'insert Module::Build'. On windows its a search and download and then copy the lib files to one of the Python @INC paths - if one runs "perl -e 'use Module::Build;'" you will see errors if this is not loaded, and you will see the path if you get an error.

2. cpan Net::Twitter

3. cpan DBI

4. write your Perl script to access the database and tweet.

The following is a very simple Perl script to test if one has competed everything as required:

(of course you will need the tokens and secrets from Twitter here)

--=====================

#!/usr/bin/perl

use Net::Twitter;

my $nt = Net::Twitter->new(

    traits              => [qw/API::RESTv1_1/],

    consumer_key        => 'xxx',

    consumer_secret     => 'yyy',

    access_token        => 'zzz',

    access_token_secret  => 'aaa',

);

$nt->update({ status => 'Hello world' });

--=====================

save this as tweet.pl and then run as 'perl tweet.pl'

And Robert is your father's brother.

Comment or email if you are doing something similar and would like help.

High level oasis diagram

To the right is a diagram of oasis' high level processes.
Currently working on the indicators (looking at log-gammas as well as optimizing current indicators), the Context and Calculate Indicators processes (putting into external C programs)

Ubuntu 13.04, Oracle 11gR2, external C programs, OCILIB

Wanting to optimize the processing and so am in the process of converting the PL/SQL for the indicator algorithms into external C programs using OCILIB.

Because I am working away from my server at the moment, and in the process of setting up ssh, I decided to put Oracle 11gR2 onto my Ubuntu 13.04 laptop. This is not really an easy process, but essentially we have the Ubuntu 12 implementation here - except we no longer need the libstdc++5 library (the link is not valid in any case, but you can find these on the Ubuntu archive site if you search for libstdc++5 if indeed you feel that you need it). This process needs to be integrated with the following document here, taking into consideration that we are installing on a deb system and not OEL 6, the important pieces are the setup of the system parameters.

Be mindful that you should install the software only first using the above guides. After implementing both guides, run 'relink all' from the $ORACLE_HOME directory. Then run netca to create the listeners and dbca to create the database itself. Then redo the Enterprise Manager having first created a password file (this is important) using emca.

The chronological order of these is ultra important otherwise you will get an 'ORA-12547 TNS lost contact' error.

Then its a question of making sure that the listener is setup correctly, this guide will help.

Also you will find that the init'dbname'.ora file is not created in the dbs directory; copy from the sample and enter the correct info, comment out the audit lines, and put in the correct Control file information.

Then OCILIB setup, and allow external program access, full database access info "host:port/service name" for the OCI_ConnectionCreate function and Robert is your father's brother.

Then its compile your programs using the following options:


gcc -Wall cprog1.c -o cprog1 -I/usr/local/include -DOCI_IMPORT_LINKAGE -DOCI_CHARSET_ANSI -locilib


This gives me is a much faster turnaround time for the rather complex and long running algorithms on PL/SQL, and allows me to play with even more complex algorithms that I have been thinking about which necessitate the use of OpenMP which I'll describe in other posts.

Comment if you need help to do something similar.

SLG, VPHA and URA

Of note, whilst we are waiting for ORE and HALO, are some current risers tweeted either yesterday or the day before:

SLG 41% today

VPHA 14% today

URA 12% today

One aspect to note from the SLG tweet is that the volume was 101 times the 3 monthly average volume, and currently today LSE is reporting 25m sells and 6m buys, with a 15% spread.

HALO

HALO

#LON:#HALO:14-MAR-13:#LSE_trace::115.9:100:9:4.38 http://t.co/MDwwsRqgSo

As you can see HALO's chart looks similar to ORE's and was also tweeted today, but with average 115.9% increase with a 7.41% spread at a current price of 27.5 and a fair cap of 7.5m GBP.

Meaning over the next 5 days we should see the price go to about 59.4, but our 11.5 % + 7.41% spread should take us to 32.7.

GLA and DYOR.

ORE

As an indicator of what I have found to be a prediction of something going on lets make a prediction here. ORE.

ORE came up on the Tweet yesterday for triggering today. The tweet reads:

#LON:#ORE:14-MAR-13:#LSE_trace::49.6:100:4:.4 http://bit.ly/XBPr8r 

and currently there are sells of 2,235,040, and only sells (at least that is what LSE is reporting, L2 shows about 50% of these are buys),  with the price holding at 0.41 with a spread of 2.5% and is a fair size at 8.8m GBP cap.

what oasis is saying is that over the next 5 days ORE should reach 0.41 plus 47.1% (on average) giving  0.60, but as I am only looking for 11.5% I am looking for 0.47 (0.41 * (11.5% required + 2.5% spread)).

Lets play.

NEOS

Would like to highlight NEOS which came up on oasis Twitter feed on 22nd Feb with a closing price of 0.105, it is currently 0.71 with an increase today of 389%.

Watching the other companies coming in on target too, although the 5 day rise is actually currently running at about 10 days, but each day at least one comes in that has been highlighted, as well as NEOS we have ARC up 16.7% just today.

I am working on the some more statistics on the rises and also tying in the success coefficients to the spread.

The Twitter feed link is in the heading of this blog, just above.

TRV

TRV did very nicely today after being triggered yesterday, up 18%

CAD

CAD is doing quite nicely today, was triggered on the 12th Feb, and today (15th Feb) currently up 29%.

Success coefficients

I have aligned the trace and candlestick success coefficients to mean the same - tweedle dee is expected percentage rise and tweedle dum is the percentage time this has occurred.

As an example below is the tweet from today for a company named CAF:

@spacecadet99999: 

#LON:#CAF:13-FEB-13:#LSE_trace::23.9:100:16: 4.13 http://t.co/MPZaFd4x

This reads as:

On the London stock exchange, for the company CAF, on 13th Feb 2013, the trace was triggered with a prior 100% success rate of obtaining 11.5% increase within the following 5 days, on the 16 occurrences this trigger happened previously, with previously the price rise averaged 23.9%, and had today a volume 4.13 times the average 3 monthly volume. The final piece is the link to the LSE website for this company.

So dee is 23.9, and dum is 100, along with the count (16) and the volume analysis (4.13) gives the map for this company. 

STK

STK from two days ago (in the legs for opera page) just went up 40% today. I will change the limits on the tweets for this type of exceptional situation.

Bid and Ask Size

It seems that trades that are dealt with by Market Makers work on the basis of Size both for Buying and Selling.

What this has meant to me is that trying to sell 100,000 BLU when the Ask Size is 70,000, the trade will not appear on L2 and that more disturbingly it may not get executed, even if the price is within limits.

Thankfully the Bid and Ask Size are on L2 and so this is another aspect of trading that really needs to be added to the whole process and which I can monitor now.

Learning something every day.

recap III

Quite a few things have changed since the last update and so lets do another one.

This is all about getting a system that I have built called 'oasis' to deliver triggers as to when to buy shares, currently geared to the London Stock Exchange as I am based in the UK, but with the ability to easily integrate other exchanges - I have run tests on NYSE, NASDAQ and FX and these worked well, so am just waiting for someone to say they would like to see these and I will reinstate. 

oasis brings down via ftp, the end of day data each evening, cleans and uploads to a database. And then the calculations start.

We have 3 trigger processes, Candlesticks, Trace and Betas.

Candlesticks

This is the usual Candlestick algorithms with added bespoke relationship algorithms. We have 8 indicators related to the relationship between periods (day).

We calculate for each company the Markov Switching Multifractal (MSM) for each individual company for these 8 Candlestick indicators. 

Each evening we compare the daily calculated indicators against the MSMs and Tweet any that have a greater than 11.5% rise indicated greater than 95% of the time.

The Tweet contains 2 specific indicators, for want of better names I have called Tweedle Dee and Tweedle Dum which are average percentage price increase and percentage times that this has risen respectively. The Tweet also contains the number of times this has triggered and the day's volume as per the average 3 monthly volume, and are labelled LSE_cs.

Trace

After End Of Day data upload all indicators are calculated (PSAR, MACD, ADX, VI, RSI, Stochastics, Bollingers and quite a few mixed combinations of these such as the PSAR on the MACD)

Each one of these indicators and combinations has been assigned an alphabetic letter for the trigger, and for every combination of 6 of these indicators we have calculated the MSM prior to the End Of Day data upload. I noticed that more than 6 indicators resulted in very limited results. Examples of this are given in this blog here.

We then compare the daily calculated indicators with the stored MSM.

As an indicator the calculation of the MSM takes about 1.5 hours per company, and takes about 4 weeks of processing to complete the whole LSE exchange companies. This process is scheduled to run constantly.

All the trigger indicators are condensed into 2 indicators again called Tweedle Dee and Tweedle Dum, which are essentially the number of indicators triggered and the number of times this has been triggered.

After daily comparisons have been made, any companies that meet the specific criteria are tweeted under LSE_trace. The Tweet also contains the number of times this has triggered and the day's volume as per the average 3 monthly volume.

Beta

This is a fully bespoke and manual set of instructions the user can enter and test. This incorporates an HTML sql interface to the database along with full reporting capabilities.

There are two servers, one production that tweets and one development that tests new algorithms.

The production server is a 16 core Xeon processor, multiple solid state disks and 32 gigs of RAM.

The development server is a 12 core Xeon with ssds and 24 gigs of RAM.

Both servers run Oracle Enterprise Linux and Oracle 11g Database.

Email or comment if you would like more information

--- 

when does the Narwhal bacon ?

Legs for opera

No Tweets today but, there is one Candlestick triggered company of note, STK, that got filtered because tweedle dum was only 84%, but the exceptional aspect was that this related to 63 occurrences with an average 36% price increase. Lets see how this gets on:

05-FEB-13 STK LSE LSE_cs uuu0202021313 63 36.06 84.13

Please note that we are currently filtering at 95% occurrence probability and so a lot fewer tweets.

Sold BLU today splitting volumes with a 28% return.

BLU

BLU is worth watching. Managed to buy in this morning prior to the 83% rise, but even though we had a sell at 11.5% the sell was not executed. Currently we have a 21.34% increase on BLU. But this does highlight the vagaries of the buy/sell process. We also did not see our sell trade on L2 and so really do not know what is going on, but would assume this to be all part of the game play. Will be experimenting with sell volumes and trying to see what the Market Makers are doing. Every step in this process seems to have its own agenda and the game is to enable some advantage to our side. Anyone have any views ? 

Candlestick analysis

Looking back at all of the Candlestick Tweets since November 2012, we have not done too badly - column definitions are in a previous post, but to clarify tweedle dee (tdee) is the average percentage increase from the Markov Fractal Chain calculations, and tweedle dum (tdum) is the percentage times this Candlestick set has been triggered and achieved the over 11.5% price increase (all are 100%):

F_CO	F_DATE	F_BETA	F_COUNT	F_TDEE	F_TDUM	VOL_	VOLANA_	CLOSE_1_	OPEN_	HIGH_5	PERC_5	PERC_10	PERC_15	PERC_30	PERC_120
AUMP	30-JAN-13	LSE_cs	3	19.33	100	79	0.01	18	14	18	28.6	28.6	28.6	28.6	28.6
BLU	24-JAN-13	LSE_cs	4	24.5	100	75000	0.83	0.3	0.25	0.3	20	20	20	20	20
JMCS	23-JAN-13	LSE_cs	3	83.17	100	1988	0.1	2.625	1.16	12.99	1019.8	1019.8	1019.8	1019.8	1019.8
BLU	15-JAN-13	LSE_cs	3	28	100	11107	0.13	0.275	0.265	0.3	13.2	13.2	13.2	13.2	13.2
SPE	14-JAN-13	LSE_cs	3	26.9	100	2109	0.05	6.625	6.55	7.25	10.7	22.1	22.1	22.1	22.1
UVEL	10-JAN-13	LSE_cs	3	19.3	100	200000	0.49	0.525	0.5025	0.55	9.5	10.4	10.4	10.4	10.4
ADS	31-DEC-12	LSE_cs	5	19.06	100	375000	0.45	0.035	0.037	0.04	8.1	8.1	21.6	21.6	21.6
BRNS	21-DEC-12	LSE_cs	3	444	100	2116	0.11	0.275	0.06	0.365	508.3	508.3	508.3	508.3	508.3
BLU	21-DEC-12	LSE_cs	3	28	100	100000	0.5	0.325	0.25	0.325	30	30	30	30	30
CRHL	21-DEC-12	LSE_cs	5	25.38	100	269019	1.13	0.35	0.3	0.35	16.7	16.7	16.7	16.7	16.7
ADS	19-DEC-12	LSE_cs	5	19.06	100	500000	0.51	0.035	0.02	0.04	100	100	125	125	125
AUMP	14-DEC-12	LSE_cs	4	23.3	100	5204	0.37	18	14	20.5	46.4	46.4	46.4	62.4	62.4
ADS	11-DEC-12	LSE_cs	5	19.06	100	150000	0.15	0.04	0.03	0.035	16.7	33.3	33.3	50	50
CMI	05-DEC-12	LSE_cs	3	50.27	100	10062	0.03	0.925	0.888	1	12.6	12.6	12.6	12.6	12.6
ATI	22-NOV-12	LSE_cs	4	17.38	100	131	0.01	71.75	62.5	74	18.4	18.4	18.4	18.4	18.4
ADS	20-NOV-12	LSE_cs	5	19.06	100	248000	0.25	0.04	0.037	0.04	8.1	8.1	8.1	21.6	21.6
FASS	13-NOV-12	LSE_cs	3	29.53	100	15431	2.4	5.005	3.52	5.01	42.3	42.3	84.7	169.9	337.2

1-Feb-2013

The following are the companies tweeted over the last couple of days on the tweet site:

https://twitter.com/spacecadet99999

F_CO	F_DATE	F_BETA	F_COUNT	F_TDEE	F_TDUM	VOL_	VOLANALYSIS_	CLOSE_1_	OPEN_	HIGH_5	PERC_5	PERC_10	PERC_15	PERC_30	PERC_120
QIL	31-JAN-13	LSE_trace	4	1000	14892	2700	3.07	0.975
AUMP	30-JAN-13	LSE_cs	3	19.33	100	79	0.01	18	14	18	28.6	28.6	28.6	28.6	28.6
YKBD	30-JAN-13	LSE_trace	4	5296	1608	320000	7.79	2.9178
MGHC	30-JAN-13	LSE_trace	3	216	1416	587873	1.73	2.4	2.74	2.74	0	0	0	0	0

No triggers or tweets today

I have been analyzing the companies that have not as yet reached close to 11.5% but were triggered in Jan 2013. These are AUR, HIW, FTD, MIRA, SIE, GBP and HVT. And of course HDY and WNER from the day before yesterday

To make it easier to follow I have added LSE links to the companies below.

F_CO	F_DATE	F_BETA	F_COUNT	F_TDEE	F_TDUM	VOL_	VOLANA_	CLOSE_1_	OPEN_	HIGH_5	PERC_5	PERC_10	PERC_15	PERC_30	PERC_120
WNER	28-JAN-13	LSE_trace	3	1034	1723	20000	0.09	1.56	1.85	1.85	0	0	0	0	0
HDY	28-JAN-13	LSE_trace	3	239	297	6058	0.12	111	108.75	109.5	0.7	0.7	0.7	0.7	0.7
BLU	24-JAN-13	LSE_cs	4	24.5	100	75000	0.83	0.3	0.25	0.3	20	20	20	20	20
JMCS	23-JAN-13	LSE_cs	3	83.17	100	1988	0.1	2.625	1.16	11.52	893.1	893.1	893.1	893.1	893.1
HVT	21-JAN-13	LSE_trace	5	2000	203	100	0.1	337.5
TND	18-JAN-13	LSE_trace	4	1955	1680	53	0.02	79.5	78.75	85	7.9	7.9	7.9	7.9	7.9
AUMP	18-JAN-13	LSE_trace	3	1063	1614	10366	1.02	18	14	22.74	62.4	62.4	62.4	62.4	62.4
VPHA	16-JAN-13	LSE_trace	4	1665	480	870242	1.67	0.735	0.735	0.75	2	8.8	8.8	8.8	8.8
BLU	15-JAN-13	LSE_cs	3	28	100	11107	0.13	0.275	0.265	0.3	13.2	13.2	13.2	13.2	13.2
SPE	14-JAN-13	LSE_cs	3	26.9	100	2109	0.05	6.625	6.55	7.25	10.7	14.5	14.5	14.5	14.5
GBP	11-JAN-13	LSE_trace	3	531	1090	1216235	6.15	8.875	9.4	9.4	0	1.1	1.1	1.1	1.1
TRP	11-JAN-13	LSE_trace	3	645	203	3921277	0.9	2.275	2.2885	2.79	21.9	21.9	21.9	21.9	21.9
UVEL	10-JAN-13	LSE_cs	3	19.3	100	200000	0.49	0.525	0.5025	0.55	9.5	9.5	9.5	9.5	9.5
SIE	10-JAN-13	LSE_trace	3	1157	1804	150182	0.51	83.36	83.575	85.025	1.7	1.7	1.7	1.7	1.7
AUMP	08-JAN-13	LSE_trace	3	1063	1614	2414	0.27	18	13.91	18	29.4	47.4	63.5	63.5	63.5
FTD	03-JAN-13	LSE_trace	3	1611	1419	2758	0.25	67.75	67	67.75	1.1	1.1	1.1	1.1	1.1
HIW	03-JAN-13	LSE_trace	3	990	990	2844048	0.62	0.89	0.929	0.929	0	0	0	0	0
WTRW	03-JAN-13	LSE_trace	3	209	580	4315	0.05	8	7.5	11.75	56.7	66.7	66.7	66.7	66.7
AUR	02-JAN-13	LSE_trace	5	3041	1855	1483115	11.33	3.625	3.8275	3.95	3.2	3.2	3.2	3.2	3.2
PCTS	02-JAN-13	LSE_trace	3	654	3340	77576	5.42	8.25	7.02	8.5	21.1	21.1	21.1	21.1	21.1
HVT	02-JAN-13	LSE_trace	5	2000	203	500	0.51	337.5	300	370	23.3	23.3	23.3	23.3	23.3